Supporting Money Management among Adults with Down Syndrome: A Multi-Technology Probe Study

Johnson, Spalitta, Kim, and Mutlu designed three cash-based budgeting technology probes for adults with Down syndrome (AwDS) and ran a comparative usage study with seven participants aged 19-39. The work is motivated by a gap between special-education financial curricula (which remain classroom-based with generic examples) and adults' real financial lives (which often consist of caregiver-managed bank accounts and small cash allowances, with guardianship laws and Supplemental Security Income rules further constraining autonomy). Each probe implemented the same three-step cash budgeting task — counting physical bills, setting budget goals, and allocating bills to envelopes — but varied the modality: (1) a gamified iPad app with drag-and-drop alien characters representing $1, $5, $10, and $20 bills; (2) an AR tablet app using Apple CreateML and a Roboflow-trained model to detect real physical bills and eliminate arithmetic; (3) a custom tangible device with two 3D-printed enclosures containing Raspberry Pi 4B computers, dual 7-inch touchscreens, magnets, tactile buttons, LEDs, a rotary knob, and a multi-level assistance switch. Sessions combined a pre-study money-handling assessment (three difficulty levels with physical U.S. currency), the three probe tasks, the User Engagement Short-Form Questionnaire, and semi-structured interviews. Data were analyzed with qualitative content analysis and interaction analysis across six consensus sessions, producing 56 descriptive codes under 12 higher-level categories.

Each modality surfaced distinct tradeoffs. The gamified app attracted attention through personalization (custom avatars and usernames) and playful characters, but its alien metaphor obscured arithmetic: participants rushed through drag-and-tap interactions, struggled to map aliens to bill denominations, and rarely used the on-screen help icon. The AR app removed arithmetic load via automatic bill detection, which helped participants with less budgeting experience — but encouraged trust in the system's output and skipping of error-checking pages, with some participants scanning bills without inspecting what had been recorded. The tangible device was the most-preferred probe (four of seven participants chose it as a favorite), with magnets, buttons, and adjustable help levels supporting engagement, though dual screens introduced coordination challenges. Prior experience shaped responses: one participant with low tech and low budgeting experience moved cautiously on the gamified app; three participants with moderate tech and allowance-only budgeting rushed and skipped checks; three with broad tech use and real financial responsibility treated the probes as tests to beat, sometimes finding AR infantilizing. Error recovery rarely emerged from interface features alone — it came from five human-support moves (reassurance, guiding questions, targeted hints, demonstrations, and direct instructions) delivered by the researcher.

For accessibility practitioners designing financial or life-skills technologies for people with intellectual and developmental disabilities, the paper delivers three concrete design implications: (1) surface budgeting as a 'stimulating multi-goal puzzle' rather than a linear sequence of steps, because AwDS can handle local numeracy tasks but struggle when choices cross-constrain each other; (2) design error recovery that explicitly connects screen state to physical cash, rather than assuming summaries or visual feedback trigger independent double-checking; (3) support interdependent use without collapsing autonomy — tools should scaffold supporter views, job-coach configurations, and user-initiated help rather than removing decisions from the AwDS. The paper also usefully contrasts gamification's attention-capture effect with the verification-skipping risk of heavy AR automation. Limitations: seven mostly White participants recruited through local community organizations, cash-only focus (no debit, bill pay, or online purchases), single-session observations that cannot speak to long-term use, and the novelty effect typical of probe studies.